Modelling of the LTDE-SD radionuclide diffusion experiment in crystalline rock at the Äspö Hard Rock Laboratory (Sweden)

This study shows a comparison and analysis of results from a modelling exercise concerning a field experiment involving the transport and retention of different radionuclide tracers in crystalline rock. This exercise was performed within the Swedish Nuclear Fuel and Waste Management Company (SKB) Task Force on Modelling of Groundwater Flow and Transport of Solutes (Task Force GWFTS). Task 9B of the Task Force GWFTS was the second subtask within Task 9 and focused on the modelling of experimental results from the Long Term Sorption Diffusion Experiment in situ tracer test. The test had been performed at a depth of about 410m in the Äspö Hard Rock Laboratory. Synthetic groundwater containing a cocktail of radionuclide tracers was circulated for 198 days on the natural surface of a fracture and in a narrow slim hole drilled in unaltered rock matrix. Overcoring of the rock after the end of the test allowed for the measurement of tracer distribution profiles in the rock from the fracture surface (A cores) and also from the slim hole (D cores). The measured tracer activities in the rock samples showed long profiles (several cm) for non-or weakly-sorbing tracers (Cl-36, Na-22), but also for many of the more strongly-sorbing radionuclides. The understanding of this unexpected feature was one of the main motivations for this modelling exercise. However, re-evaluation and revision of the data during the course of Task 9B provided evidence that the anomalous long tails at low activities for strongly sorbing tracers were artefacts due to cross-contamination during rock sample preparation. A few data points remained for Cs-137, Ba-133, Ni-63 and Cd-109, but most measurements at long distances from the tracer source (>10mm) were now below the reported detection limits. Ten different modelling teams provided results for this exercise, using different concepts and codes. The tracers that were finally considered were Na-22, Cl-36, Co-57, Ni-63, Ba-133, Cs-137, Cd-109, Ra-226 and Np-237. Three main types of models were used: i) analytical solutions to the transport-retention equations, ii) continuum-porous-medium numerical models, and iii) microstructure-based models accounting for small-scale heterogeneity (i.e. mineral grains, porosities and/or microfracture distributions) and potential centimetre-scale fractures. The modelling by the different teams led to some important conclusions, concerning for instance the presence of a disturbed zone (a few mm in thickness) next to the fracture surface and to the wall of the slim hole and the role of micro-fractures and cm-scale fractures in the transport of weakly sorbing tracers. These conclusions could be reached after the re-evaluation and revision of the experimental data (tracer profiles in the rock) and the analysis of the different sets of model results provided by the different teams.The comments from Dr. Kersti Nilsson, the analytical work by VKTA (Dresden, Germany) for some of the rock samples, and the initial review by Dr. Anna-Maria Jakobsson are gratefully acknowledged. The constructive comments and suggestions by Dr. Jordi Cama and an anonymous reviewer contributed to a significant improvement of the manuscript. Funding was provided through the Task Force partner organisations participating in this modelling exercise (SKB, Sweden; POSIVA OY, Finland; SÚRAO, Czech Republic; BMWi, Germany; KAERI, Republic of Korea; NUMO and JAEA, Japan). IDAEA-CSIC is a Severo Ochoa Centre of Research Excellence (Spanish Ministry of Science and Innovation, Project CEX2018-000794-S). The contributions of TUL, CTU and PROGEO are the result of the SÚRAO project “Research support for Safety Evaluation of Deep Geological Repository”. JAEA’s modelling was performed as a part of “The project for validating assessment methodology in geological disposal system” funded by the METI of Japan. A.I., P.T., M.V., G.D., and D.B. gratefully acknowledge the computing time granted by the JARA Vergabegremium and provided on the JARA Partition part of the supercomputer JURECA at Forschungszentrum Jülich.Peer reviewe